DAOS as HPC Storage, a view from Numerical Weather Prediction

Nicolau Manubens Gil; Tiago Quintino; Simon D. Smart; Emanuele Danovaro; William Adrian Jackson

doi:10.1109/IPDPS54959.2023.00106

DAOS as HPC Storage, a view from Numerical Weather Prediction

Nicolau Manubens Gil, Tiago Quintino, Simon D. Smart, Emanuele Danovaro, William Adrian Jackson

Research output: Contribution to conference › Paper › peer-review

Abstract / Description of output

Object storage solutions potentially address long-standing performance issues with POSIX file systems for certain I/O workloads, and new storage technologies offer promising performance characteristics for data-intensive use cases.

In this work, we present a preliminary assessment of Intel's Distributed Asynchronous Object Store (DAOS), an emerging high-performance object store, in conjunction with non-volatile storage and evaluate its potential use for HPC storage. We demonstrate DAOS can provide the required performance, with bandwidth scaling linearly with additional DAOS server nodes in most cases, although choices in configuration and application design can impact achievable bandwidth. We describe a new I/O benchmark and associated metrics that address object storage performance from application-derived workloads.

Original language	English
Pages	1029-1040
Number of pages	12
DOIs	https://doi.org/10.1109/IPDPS54959.2023.00106
Publication status	Published - 2023
Event	37th IEEE International Parallel and Distributed Processing Symposium - St Petersburg, United States Duration: 15 May 2023 → 19 May 2023 Conference number: 37

Conference

Conference	37th IEEE International Parallel and Distributed Processing Symposium
Abbreviated title	IPDPS
Country/Territory	United States
City	St Petersburg
Period	15/05/23 → 19/05/23

Keywords / Materials (for Non-textual outputs)

scalable object storage
next-generation I/O
Storage class memory
non-volatile memory
Numerical weather prediction
DAOS

Access to Document

10.1109/IPDPS54959.2023.00106Licence: All Rights Reserved

https://arxiv.org/abs/2208.06752Licence: Creative Commons: Attribution (CC-BY)

Extreme-scale precision Imaging in Radio Astronomy (EIRA)
Jackson, A.
EPSRC
1/09/20 → 31/08/24
Project: Research

Cite this

@conference{53901108dafc4f10b6df1a13f493ac8f,

title = "DAOS as HPC Storage, a view from Numerical Weather Prediction",

abstract = "Object storage solutions potentially address long-standing performance issues with POSIX file systems for certain I/O workloads, and new storage technologies offer promising performance characteristics for data-intensive use cases. In this work, we present a preliminary assessment of Intel's Distributed Asynchronous Object Store (DAOS), an emerging high-performance object store, in conjunction with non-volatile storage and evaluate its potential use for HPC storage. We demonstrate DAOS can provide the required performance, with bandwidth scaling linearly with additional DAOS server nodes in most cases, although choices in configuration and application design can impact achievable bandwidth. We describe a new I/O benchmark and associated metrics that address object storage performance from application-derived workloads.",

keywords = "scalable object storage, next-generation I/O, Storage class memory, non-volatile memory, Numerical weather prediction, DAOS",

author = "{Manubens Gil}, Nicolau and Tiago Quintino and Smart, {Simon D.} and Emanuele Danovaro and Jackson, {William Adrian}",

note = "Funding Information: This work has also been contributed by the the ACROSS and IO-SEA projects, funded by the European High-Performance Computing Joint Undertaking (JU) under Grant Agreements no. 955648 and no. 955811, respectively. The JU receives support from the European Union's Horizon 2020 research and innovation programme and Italy, France, Czech Republic, United Kingdom, Greece, Netherlands, Germany, Norway. The NEXTGenIO system was funded by the European Union's Horizon 2020 Research and Innovation program under Grant Agreement no. 671951, and supported by EPCC, The University of Edinburgh. For the purpose of open access, Adrian Jackson has applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising from this submission. Funding Information: The NEXTGenIO system was funded by the European Union{\textquoteright}s Horizon 2020 Research and Innovationprogram under Grant Agreement no. 671951, and supported by EPCC, The University of Edinburgh. Funding Information: This work has also been contributed by the the ACROSS and IO-SEA projects, funded by the European High-Performance Computing Joint Undertaking (JU) under Grant Agreements no. 955648 and no. 955811, respectively. The JU receives support from the European Union{\textquoteright}s Horizon 2020 research and innovation programme and Italy, France, Czech Republic, United Kingdom, Greece, Netherlands, Germany, Norway. Funding Information: ACKNOWLEDGMENT The work presented in this paper was carried out with funding by the European Union under the Destination Earth initiative (cost center DE3100, code 3320) and relates to tasks entrusted by the European Union to the European Centre for Medium-Range Weather Forecasts. Views and opinions expressed are those of the author(s) only and do not necessarily reflect those of the European Union or the European Commission. Neither the European Union nor the European Commission can be held responsible for them. Publisher Copyright: {\textcopyright} 2023 IEEE.; 37th IEEE International Parallel and Distributed Processing Symposium, IPDPS ; Conference date: 15-05-2023 Through 19-05-2023",

year = "2023",

doi = "10.1109/IPDPS54959.2023.00106",

language = "English",

pages = "1029--1040",

}

TY - CONF

T1 - DAOS as HPC Storage, a view from Numerical Weather Prediction

AU - Manubens Gil, Nicolau

AU - Quintino, Tiago

AU - Smart, Simon D.

AU - Danovaro, Emanuele

AU - Jackson, William Adrian

N1 - Conference code: 37

PY - 2023

Y1 - 2023

N2 - Object storage solutions potentially address long-standing performance issues with POSIX file systems for certain I/O workloads, and new storage technologies offer promising performance characteristics for data-intensive use cases. In this work, we present a preliminary assessment of Intel's Distributed Asynchronous Object Store (DAOS), an emerging high-performance object store, in conjunction with non-volatile storage and evaluate its potential use for HPC storage. We demonstrate DAOS can provide the required performance, with bandwidth scaling linearly with additional DAOS server nodes in most cases, although choices in configuration and application design can impact achievable bandwidth. We describe a new I/O benchmark and associated metrics that address object storage performance from application-derived workloads.

AB - Object storage solutions potentially address long-standing performance issues with POSIX file systems for certain I/O workloads, and new storage technologies offer promising performance characteristics for data-intensive use cases. In this work, we present a preliminary assessment of Intel's Distributed Asynchronous Object Store (DAOS), an emerging high-performance object store, in conjunction with non-volatile storage and evaluate its potential use for HPC storage. We demonstrate DAOS can provide the required performance, with bandwidth scaling linearly with additional DAOS server nodes in most cases, although choices in configuration and application design can impact achievable bandwidth. We describe a new I/O benchmark and associated metrics that address object storage performance from application-derived workloads.

KW - scalable object storage

KW - next-generation I/O

KW - Storage class memory

KW - non-volatile memory

KW - Numerical weather prediction

KW - DAOS

U2 - 10.1109/IPDPS54959.2023.00106

DO - 10.1109/IPDPS54959.2023.00106

M3 - Paper

SP - 1029

EP - 1040

T2 - 37th IEEE International Parallel and Distributed Processing Symposium

Y2 - 15 May 2023 through 19 May 2023

ER -

DAOS as HPC Storage, a view from Numerical Weather Prediction

Abstract / Description of output

Conference

Keywords / Materials (for Non-textual outputs)

Access to Document

Fingerprint

Projects

Extreme-scale precision Imaging in Radio Astronomy (EIRA)

Cite this